Peter Cotton was one of the architects at the heart of the CDO market but got out before the wheels came off the wagon. He left Morgan Stanley in 2007 to form venture-funded Julius Finance, a "model-fusion specialist" which promises "unprecedented visibility for market derived pricing and analytics" for the $58 trillion credit derivatives market.
Cotton, equipped with a PhD in math from Stanford University, believes his better mousetrap can get things back on track.
BTN: You left Morgan Stanley's structured credit products group in 2007. Did you see this coming?
Peter Cotton: (laughs) Um, no, is the short answer. I did not see this coming. Well, I was working on the credit risk models, and they're a bit like seat belts: I knew they didn't work, but I didn't see the crash coming.
BTN: What from what you saw there made you realize the models didn't work?
Cotton: Well, I built them. The demands of the trading floor mean that it's possible to build models - really interpolation widgets - to mark the books. But building things that can be used to simulate the future is much harder, and it's very difficult to get the time and the resources to do that.
BTN: You've called for "banker proof" CDO valuations. What do you mean?
Cotton: If you only have tools that provide local interpolations and which don't actually generate scenarios, then it's very easy to game the system because the models are all internally inconsistent and you can kind of go from A to B then from B to C and C to A and not get back where you started. So there's all sorts of games that can be played, in addition to the usual rating agency games. We worry that the government is going to become an enormous customer of structured credit products and doesn't really have the expertise to do so, and is likely to be fooled the same way that clients were in the past.
BTN: Does anyone have the expertise?
Cotton: There are very few research and development efforts on Wall Street. It was all very short-term work and this hasn't really left the industry in a very good place to deal with this crisis.
Why focus on credit derivatives?
BTN: The reason for focusing on the credit derivatives first is the modeling there was particularly bad, compared with say, interest rate markets, where the modeling is actually quite good.
What about the models that got us here can they be salvaged or do institutions have to start over?
Cotton: I think to a large extent institutions need to start over. Most financial institutions didn't have very much in the way of modeling at all, they relied very heavily on the ratings agencies. The existing kind of platforms can be quite sophisticated and sort of shiny on the outside but at the core they typically contain some very poor assumptions.
BTN: What about the future of financial modeling?
Cotton: I think the public perception is that these complicated models created these complicated products and got us into this mess - it's all the quants fault!...But anybody whose worked on a trading floor knows the reality is very different and the quants more or less had a choice of doing what they're told or being pushed out the door. I think that the lesson in this is there's actually a difference between getting it right and getting it wrong.